Why it is different
Verifiable privacy
The gateway runs in a TEE and proves it with a hardware-backed
attestation report. It does not retain request or response
bodies.
Confidential models
Confidential models run on verified TEE providers, with the channel
bound before your prompt is forwarded.
Drop-in compatible
Point your OpenAI or Anthropic SDK at the base URL. Same request and response shapes you already
use.
Pick your path
Build with the API
Point your OpenAI or Anthropic SDK at one endpoint. Confidential and standard models, with
attestation and a signed receipt on every private request.
Use in agents
Run Claude Code, Codex, Pi, OpenCode, and more through Private AI Proxy, a desktop app that
verifies the gateway and connects your agents.
Use the chat app
End-to-end encrypted chat with confidential models and private document search.
Verify, don’t trust
You do not have to take the API’s word. After one request you can check, locally, that:- The gateway is a measured workload running in a genuine TEE. Pin a reviewed compose hash when you also need release acceptance. Attestation report →
- The response came from that workload, was not altered, and (for a confidential model) ran on a verified upstream enclave. Receipts →
Start building
1
Get an API key
Create a key in the dashboard and store it as an environment
variable. See Authentication.
2
Make your first call
Point an OpenAI client library at
https://api.redpill.ai/v1. See the
Quickstart.3
Verify the response
Use an ACI client to verify the attestation, pinned TLS channel, receipt, and upstream session.
Verify a response.